It's just another type of report, and
I think we were wise to choose xslt as our preferred technology for
reports. Using xslt, we easily get all the flexibility you mention.
I see no reason to treat this report otherwise.
This is different from other reports in a few ways:
(1) It would be run at the "case" level. To our salespeople, a
"case" is a set of individuals who have similar insurance policies
and are related to each other in some way--typically, they'd be
employees of the same employer. In lmi, a '.cns' file embodies a
"case". It would make no sense to produce this report for a single
individual. The command to produce it would appear on the "Census"
menu that's available when a '.cns' file is loaded.
(2) It is not required by law or regulation, and therefore it
wouldn't need xsl-fo.
(3) It may use some data that are not currently available in our
largest "master" data set, which therefore may need to be expanded.
We'll have to study the problem more to know that for sure. That's
not a serious obstacle: the "ledger" objects that provide the data
we use for all reports already contain most of what we'd need, and
any new data that we might need to add would be similar to what
we've already got.